Brain Tumor Detection from Multimodal MRI Brain Images using Pseudo Coloring Processes

Abstract

Abstract In this work, we come up with a novel method to merge the color channels created from multi modal images to form a RGB image and then highlight the tumor in the MRI brain images, detect and extract the tumor region from the human head scan MRI images. This extraction process is done purely based on multimedia color channels and the basic characteristics of a MRI brain image. The original multi modal grayscale images (FLAIR, T2, T1c) are used to build the Red, Green and Blue channels respectively and then combined to form a whole RGB image using the pseudo coloring process. The Pseudo colored brain images are then analyzed for the presence of specific colors in the images which clearly identifies the tumor region in the whole brain¬ image. The tumor and necrotic regions are shown in the RGB images in Golden Yellow color and White color respectively. Experimental study was done on BRATS high glioma imagesets. The main objective of this work is to detect the tumor in the human brain scans. A detailed study has to be done to detect the various regions of tumor. The results given by the proposed method reached good qualitative and quantitative level.